Completion norms for 3085 English sentence contexts.

Area of Science:

• Psycholinguistics
• Cognitive Science
• Computational Linguistics

Background:

• Sentence context significantly influences word processing in language comprehension.
• Understanding word predictability is crucial for modeling reading and listening.
• Existing resources for word predictability norms are limited.

Purpose of the Study:

• To create a comprehensive dataset of sentence completion norms for English.
• To provide quantitative measures of word predictability based on context.
• To support research in psycholinguistics and natural language processing.

Main Methods:

• Collected sentence completion data online using a written cloze procedure.
• Utilized 3,085 English sentences (8-10 words long) with at least 100 unique participants per sentence.
• Human raters reviewed responses to ensure data quality, mitigating errors.

Main Results:

• Generated predictability values for 13,438 unique target words.
• Identified 6,790 words appearing in multiple sentence contexts.
• Calculated entropy values reflecting the distribution of word responses.

Conclusions:

• The developed norms offer a valuable resource for studying language processing.
• The data facilitates research on how context shapes word prediction.
• Accessible norms and code support further investigations in the field.